import os
import pandas as pd
import numpy as np
import plotly.express as px
import seaborn as sns
import matplotlib.pyplot as plt
from scipy.stats import ttest_ind, skew, kurtosis

# Load the cleaned dataset
merged_data_filename = os.path.join("..", "data", "processed", "Credit_Card_Financial.csv")
merged_df = pd.read_csv(merged_data_filename)

print(f"Dataset Shape: {merged_df.shape}")
print(f"Total Customers: {merged_df['Client_Num'].nunique()}")

# Check for missing values
missing_values = merged_df.isnull().sum()
print("Missing Values Summary:")
print(missing_values[missing_values > 0] if missing_values.sum() > 0 else "No missing values found")

# Basic dataset info
print(f"\nDataset Info:")
print(f"Shape: {merged_df.shape}")
print(f"Memory Usage: {merged_df.memory_usage(deep=True).sum() / 1024**2:.2f} MB")

# Calculate correlations for key financial variables
corr_vars = ['Income', 'Credit_Limit', 'Total_Trans_Amt', 'Annual_Fees']
corr_matrix = merged_df[corr_vars].corr()

print("Key Correlation Coefficients:")
print(f"Income vs Total_Trans_Amt: {corr_matrix.loc['Income', 'Total_Trans_Amt']:.3f}")
print(f"Income vs Credit_Limit: {corr_matrix.loc['Income', 'Credit_Limit']:.3f}")
print(f"Credit_Limit vs Total_Trans_Amt: {corr_matrix.loc['Credit_Limit', 'Total_Trans_Amt']:.3f}")

plt.figure(figsize=(5, 4))
sns.heatmap(corr_matrix, annot=True, cmap='Blues', center=0, 
            square=True, linewidths=0.5, cbar_kws={"shrink": .8})
plt.title("Correlation Matrix of Key Financial Variables")
plt.tight_layout()
plt.show()

# Calculate skewness and kurtosis
credit_limit_skew = skew(merged_df["Credit_Limit"].dropna())
credit_limit_kurt = kurtosis(merged_df["Credit_Limit"].dropna())

print("Credit Limit Distribution Analysis:")
print(f"Skewness: {credit_limit_skew:.3f} (Right-skewed if > 0)")
print(f"Kurtosis: {credit_limit_kurt:.3f} (Heavy tails if > 0)")
print(f"Mean: ${merged_df['Credit_Limit'].mean():,.0f}")
print(f"Median: ${merged_df['Credit_Limit'].median():,.0f}")

gender_dist = merged_df['Gender'].value_counts().reset_index()
gender_dist.columns = ['Gender', 'Count']

fig = px.pie(gender_dist, names='Gender', values='Count', 
             title="Gender Distribution",
             color_discrete_sequence=["#003399", "#3366cc"])

fig.update_traces(texttemplate="%{percent:.1%}", textposition="inside")
fig.update_layout(showlegend=True, margin=dict(l=60, r=50, t=50, b=150))
fig.show()

income_by_education = merged_df.groupby('Education_Level')['Income'].mean().reset_index()
income_by_education = income_by_education.sort_values('Income', ascending=False)

fig = px.bar(income_by_education, x='Education_Level', y='Income',
             title='Average Income by Education Level',
             color_discrete_sequence=['#002366'])

fig.update_layout(xaxis_title="Education Level", yaxis_title="Average Income ($)", 
                 bargap=0.4, bargroupgap=0.2, margin=dict(l=60, r=50, t=50, b=150))
fig.show()

credit_by_card = merged_df.groupby('Card_Category')['Credit_Limit'].mean().reset_index()
credit_by_card = credit_by_card.sort_values('Credit_Limit', ascending=False)

fig = px.bar(credit_by_card, x='Card_Category', y='Credit_Limit',
             title='Average Credit Limit by Card Type',
             color_discrete_sequence=['#002366'])

fig.update_layout(xaxis_title="Card Type", yaxis_title="Average Credit Limit ($)", 
                 bargap=0.4, bargroupgap=0.2, margin=dict(l=60, r=50, t=50, b=150))
fig.show()

# Perform t-test
men_credit = merged_df[merged_df['Gender'] == 'Male']['Credit_Limit']
women_credit = merged_df[merged_df['Gender'] == 'Female']['Credit_Limit']

t_stat, p_value = ttest_ind(men_credit, women_credit, nan_policy='omit')

print("T-Test Results: Gender vs Credit Limit")
print(f"T-statistic: {t_stat:.3f}")
print(f"P-value: {p_value:.6f}")
print(f"Significant difference: {'Yes' if p_value < 0.05 else 'No'}")
print(f"Male average Credit Limit: ${men_credit.mean():,.0f}")
print(f"Female average Credit Limit: ${women_credit.mean():,.0f}")

# Create month order for proper sorting
month_order = ['January', 'February', 'March', 'April', 'May', 'June',
               'July', 'August', 'September', 'October', 'November', 'December']

# Group by month and calculate total transactions
monthly_trans = merged_df.groupby('Month')['Total_Trans_Amt'].sum().reset_index()
monthly_trans['Month'] = pd.Categorical(monthly_trans['Month'], categories=month_order, ordered=True)
monthly_trans = monthly_trans.sort_values('Month')

fig = px.line(monthly_trans, x='Month', y='Total_Trans_Amt',
              title='Total Transaction Amount Over Time',
              markers=True)

fig.update_traces(line_shape='spline', line=dict(color='#002366', width=3))
fig.update_layout(xaxis_title="Month", yaxis_title="Total Transaction Amount ($)", 
                 margin=dict(l=80, r=30, t=50, b=150))
fig.show()

delinq_by_state = merged_df.groupby('state_cd')['Delinquent_Acc'].sum().reset_index()
delinq_by_state = delinq_by_state.sort_values('Delinquent_Acc', ascending=False).head(10)

fig = px.bar(delinq_by_state, x='Delinquent_Acc', y='state_cd',
             orientation='h',
             title='Top 10 States with Highest Delinquent Accounts',
             color_discrete_sequence=['#002366'])

fig.update_layout(xaxis_title="Number of Delinquent Accounts", yaxis_title="State", 
                 margin=dict(l=150, r=50, t=50, b=150))
fig.show()

satisfaction_dist = merged_df.groupby('Cust_Satisfaction_Score')['Client_Num'].count().reset_index()
satisfaction_dist.columns = ['Satisfaction_Score', 'Customer_Count']

fig = px.bar(satisfaction_dist, x='Satisfaction_Score', y='Customer_Count',
             title='Customer Count by Satisfaction Level',
             color_discrete_sequence=['#002366'])

fig.update_layout(xaxis_title="Satisfaction Score", yaxis_title="Number of Customers", 
                 bargap=0.4, bargroupgap=0.2, margin=dict(l=80, r=50, t=50, b=150))
fig.show()